多智能体模拟在规划师知识构建及空间规划决策支持中的应用——以日本地方城市老年人日护理中心空间战略规划为例

发布时间：2018-03-17 07:18

本文选题：多智能体模型　切入点：规划支持系统　出处：《现代城市研究》2016年11期 　论文类型：期刊论文

【摘要】：面向城市问题的复杂性和动态性,大量模拟仿真、定量分析等手段被广泛应用于城乡规划方案的预测与评价中,而其相应的预测评价结果也往往被认为有助于规划方案的调整与决策。在此过程中,仿真模型的构建逐渐从宏观向微观方向发展,对研究对象的分析也逐渐体现出从群体性描述向个体行为特征、机理的刻画转变。但是这其中却鲜有学者探讨基于微观个体行为的仿真模拟究竟在规划方案的生成与决策过程中发挥怎样的作用,对规划师的知识构建是否能够有效地达到规划决策支持的目的。本研究从这一科学问题出发,基于一个面向养老日护理中心空间分布模拟的多智能体模型(Agent-Based Model,ABM)的构建与应用,探讨多智能体模拟仿真在空间规划实践中的应用可能及其构建机制。研究结果表明,多智能体模拟用于规划实践首先必须具有可信的、完整的数据支持;要达到辅助规划师建立对规划问题的科学认识,模型构建则应该在规划工作的最初开始介入,并让规划师参与其中;同时,多智能体模型的构建必须基于面向规划问题的现实决策过程。总体而言,多智能体模型可用于规划决策支持,但是模拟结果却不能代替规划方案。
[Abstract]:Facing the complexity and dynamics of urban problems, a large number of simulation, quantitative analysis and other means are widely used in the prediction and evaluation of urban and rural planning plans. The corresponding prediction and evaluation results are often considered to be helpful to the adjustment and decision of the planning scheme. In this process, the construction of simulation model gradually develops from macro to micro direction. The analysis of the object of study gradually reflects the characteristics of individual behavior from group description to individual behavior. However, few scholars have discussed how simulation based on micro-individual behavior plays a role in the process of planning scheme generation and decision-making. Whether the knowledge construction of planners can effectively achieve the purpose of planning decision support, this study based on the construction and application of Agent-Based Modeler (ABM), a multi-agent model for spatial distribution simulation of care centers for the aged, is carried out in this paper. The application of multi-agent simulation in spatial planning practice and its construction mechanism are discussed. The results show that multi-agent simulation for planning practice must first have credible and complete data support. To assist planners in building a scientific understanding of planning issues, model building should begin at the very beginning of the planning process and involve the planners; at the same time, In general, the multi-agent model can be used for planning decision support, but the simulation results can not replace the planning scheme.
【作者单位】：日本金泽大学;福州大学建筑学院;
【分类号】：TU984.113

【相似文献】